HWRF Surface Layer Thermodynamics Evaluation. Eric W. Uhlhorn and Joseph J. Cione HFIP Hurricane Modeling Workshop September 2012

Transcription

1 HWRF Surface Layer Thermodynamics Evaluation Eric W. Uhlhorn and Joseph J. Cione HFIP Hurricane Modeling Workshop September 2012

2 Special thanks. HRD HWRF modeling team Gopal, Xuejin Zhang, Thiago Quirino, Vijay Tallapragada HFIP

3 Hurricane Air-Sea Interaction Physics

4 Evaluation of coupled air-sea thermodynamics Observations from Tropical Cyclone Buoy Database (TCBD; Cione et al. 2000, Cione and Uhlhorn 2003, Cione et al. in review) HWRF 2011 retrospective model runs GPS dropwinsonde database (Zhang et al. 2011, Zhang and Uhlhorn 2012, Zhang et al. in preparation) Consider observations in hurricanes only Radially between 0.5 (no eye) and 6 RMWs

5 TC Buoy Observations Cione et al. 2000, 2003 Temperature and humidity reported hourly Winds (10-min mean) reported every 10 mins. Obs. adjusted to 10-m level Winds converted to 1-min mean

6 2011 Season HWRF Retrospective Runs Configuration 3 km inner nest Coupled to ocean (POM) Modified C k, C d (CBLAST & others) Modified diffusion (Zhang et al. 2011) Operational in 2012 From Gopalakrisnan et al (JAS, in press)

7 2011 Season HWRF Retrospective Runs (cont.) Storms Irene-09L (34 runs) Katia-11L (46 runs) Maria-14L (41 runs) Ophelia-16L (48 runs) Philippe-17L (60 runs) Rina-18L (20 runs) Total 249 runs, 126 hr simulations, output every 3 hours

8 2011 HWRF & TCBD Storm Stats TCBD HWRF

9 Model Sampling Methodology Fields of SST, latent (Q l ) and sensible (Q s ) heat fluxes, and 10-m winds (U 10 ) are provided as model output T q C h SST q s C e C c k p ( SST ) Q C s h U Ql L C U v 10 e 10 Compute T 10, q 10 from output model fields Sample model at TCBD buoy locations falling within model grid as cyclones translate/evolve Compute statistical distributions and compare with observations

10 Radial Distributions of Observed and HWRF Surface Layer SST T 10 q 10 DT Dq

11 Observed vs. Simulated SST Variability TCBD HWRF

12 Hurricane Katia Buoy Intercept

13 GPS Dropwindsonde Database 10 m winds, temperature, humidity, NO SST Ongoing (labor-intensive) effort to add co-located AXBT SSTs to sonde profiles

14 Empirical Probability Distributions

15 Observed Temperature and Moisture Relative Humidity Temperature and Dewpoint

16 HWRF Evaluation

17 Controls on Moisture Flux TCBD Obs. q 10 better correlated with Dq than q s Both de-correlate slightly with increased wind, but q 10 remains more highly correlated HWRF q s better correlated with Dq than q 10 q 10 shows almost no relationship to to Dq at high winds

18 Summary A comprehensive evaluation of air-sea thermodynamic properties of the operational coupled HWRF has been performed. Results indicate: HWRF atmosphere near-surface is typically warmer and more moist than observed HWRF surface layer significantly lower relative humidity Gradual tendency toward saturation as wind speed increases (max ~97% at 60 m/s) Obs show far more rapid trend toward saturation (>95% at 30 m/s, max ~97-98% at 40 m/s) POM-simulated SST cools significant less than observed in response to TC forcing

19 Questions How to eliminate warm SST bias? URI colleagues have indicated wind stress is reduced 25% in operational version for 2012 Comprehensive observation-based evaluation of operational coupled POM needed How to cool/dry the surface layer? Fluxes and exchange coefficient modifications Spray (cools but moistens) Precipitation-induced downdraft transport Entrainment/shallow convection Is HWRF overly-sensitive to the ocean? Coupled-model simulations have shown changes to the ocean coupling can have large impact on simulated intensity Extensive coupled-modeling efforts have not led to significant improvements to intensity prediction